AX: The Next Evolution in AI Agent Software Design
Mathias Biilmann, CEO of Netlify, explores how Agent Experience (AX) will become as crucial as UX and DX in software development.
Link, Description & Synopsis
Link:
Introducing AX: Why Agent Experience Matters
Description:
Mathias Biilmann, CEO of Netlify, explores how Agent Experience (AX) will become as crucial as UX and DX in software development.
Synopsis:
This article covers how to:
Design software specifically for AI agent interaction
Create open platforms that support multiple agent types
Balance human and agent needs in software design
Implement practical AX improvements in existing platforms
Context
As AI agents infiltrate the economy, software must evolve to effectively support human and agent users.
The evolution of software design has seen three major paradigm shifts.
In 1993, Don Norman coined “user experience” (UX) to describe how humans interact with systems.
UX brought human-centered design to the forefront of software design.
By 2011, as software became more platform-oriented, Jeremiah Lee introduced “developer experience” (DX) to capture how developers build upon existing systems.
DX recognized developers as a distinct user category with unique needs.
In 2025, agent experience (AX) emerges as the third major paradigm in this evolution.
AX now acknowledges AI agents as first-class users of our systems.
Each transition marked a fundamental shift in how we think about software design.
The timing is particularly relevant as organizations grapple with integrating AI capabilities into their platforms.
Key Implementation Patterns
The article outlines several emerging patterns in AX design:
Platform Integration: Infrastructure
Clean, well-described APIs for agent interaction
Machine-readable documentation
Streamlined authentication flows
Clear versioning and compatibility guidelines
Platform Integration: Open Design
Support for multiple agent types
User choice in agent selection
Collaborative workflows between humans and agents
Platform-agnostic integration capabilities
Platform Integration: Agent-First Architecture
Purpose-built API endpoints for agent use
Automated deployment pipelines
Resource optimization for agent workflows
Built-in verification and validation
These patterns reflect a growing understanding that AI agents require different design considerations than human users or developers.
Strategic Implications
The article points to several crucial strategic considerations:
Development Focus
API design optimized for agent consumption
Documentation tailored for machine reading
Integration patterns for multiple agent types
Balance between automation and human oversight
Market Evolution
Movement from closed to open-agent ecosystems
Competitive advantage through agent accessibility
Growing importance of platform interoperability
Focus on agent-human collaboration
Business Impact
Increased development velocity through agent collaboration
Lower barriers to custom development
Exponential growth in web applications
New paradigms in software interaction
These strategic considerations suggest that organizations need a clear framework for implementing AX in their platforms.
Implementation Framework
For organizations implementing AX:
Platform Assessment
Evaluate current agent accessibility
Identify integration bottlenecks
Map potential agent interactions
Define success metrics
Infrastructure Updates
Enhance API documentation
Implement agent authentication
Build verification systems
Create monitoring tools
Ecosystem Development
Foster agent partnerships
Build developer communities
Create example implementations
Share best practices
As teams implement these frameworks, several key insights emerge about the future of software design.
Personal Notes
The emphasis on open ecosystems rather than closed, proprietary agent systems is exciting.
This emphasis on open ecosystems mirrors the evolution of the web itself, where open standards and interoperability have driven and continue to drive innovation and growth.
Looking Forward: Agent Integration
The industry will likely evolve to include:
Standardized agent interaction protocols
Cross-platform agent authentication
Unified documentation standards
Collaborative development environments
Advanced agent-human workflows
Conclusion
The shift toward AX represents a fundamental change in how we think about software design, moving from human-only interfaces to systems that seamlessly support human and agent users.